Research On Key Technologies Of Path Planning For Underwater Vehicle

The exploration of the marine resources has become an "invisible" science and te chnology competition around the world.It plays an important role of the underwater v ehicle path planning technology in the exploration of the ocean.With the increasing d emand for the exploration of marine resources of human and the increasing complexit y of the seabed environment,the multi-underwater vehicle system came into being bec ause the single underwater vehicle can no longer meet the requirements.This system has the advantages of strong maneuverability,high efficiency and better response to e mergencies.On this basis,there will be a great improvement of the whole system by adding a multiple underwater robot.The efficiency of the whole system will be greatl y improved if the multi-vehicle synergy is added.The key technology of path plannin g for underwater vehicle was deeply studied in this paper.The 3d modeling emulation technique of underwater environment,global path planning technique of multiple unde rwater vehicles,real-time collision avoidance technology of underwater dynamic obstacl e,strong target tracking technology of complex environment,and cooperative localizati on technology of multiple underwater vehicles were included in this study.Main conte nts as follows:Submarine environment modeling is the premise of underwater vehicle path planni ng.Accurately establish the submarine topography can increase the safety index and i mprove the efficiency of underwater navigation.Submarine environment modeling is th e premise of underwater vehicle path planning,and the accurate establishment of seab ed topography can increase the safety index of underwater vehicle navigation and impr ove the navigation efficiency.Discrete depth data points are extracted from multiple t wo-dimensional vector electronic charts in this paper.Firstly,several classical data fusi on methods are introduced,and then explain their advantages and disadvantages.Consi dering the complexity and low efficiency of terrain data fusion,an improved self-corre ction fusion technology was proposed by combining the intelligent optimization algorit hm and kalman filter.and the discrete water depth points were obtained by fusion.W e fuse the extracted discrete water depth points.The integrated depth point has a high utilization efficiency,while the relatively large interval makes it impossible to simulat e the seabed environment with high accuracy.Based on the theory of fractal interpolat ion algorithm,we interpolates the fused data and proposes a compound fractal interpol ation algorithm.A more accurate three-digit seabed environment model generated.Fina lly,the improved interpolation algorithm is compared with the traditional interpolation algorithm by numerical simulation.Path planning is divided into static planning and dynamic planning.The global st atic path planning problem is transformed into a evaluation problem of segmented opti mal point in this paper.The inertia weights and learning factors of particle swarm opt imization(pso)are improved to avoid falling into "precocity".Considering that particle swarm optimization algorithm has the characteristics of memory and parallel computin g,it is used for global path planning of two underwater vehicles.In view of the und erwater vehicle cannot carry on the sharp turn and the flexibility is weak,we propose s a two-parameter smooth subdivision algorithm.The algorithm is used to smooth the initial path,and the convergence of this method is proved in detail.The improved alg orithm is used to simulate the global static path planning of two underwater vehicles,and experimental results are described analytically in the end.In view of the uncertain factors in dynamic path planning,we mainly consider th e influence of dynamic large fish and large moving floating objects.A dynamic obstac le avoidance strategy based on relative motion model is proposed by studying the kine matics principle of underwater vehicle and dynamic obstacle.The visual sensor and so nar sensor are used to acquire the data of the near and long distance dynamic obstacl e respectively,and the trajectory of obstacles is predicted combined with extended kal man filter.At the same time,the collision avoidance strategies for speed reduction an d turning processing are established.Different weights were set for collision avoidance strategies of different models.Then the method is transformed into integer linear opti mization problem and numerical simulation experiments are carried out for this metho d.Nowadays,the path planning technology is no longer a collision-free trajectory pr oblem in the traditional sense.It needs to meet the actual demand of engineering appl ication and add target strong tracking technology on this basis.While the strong target tracking technique requires a filtering model with strong robustness and high stability.To solve this problem,we puts forward a strong robustness H? filtering algorithm w hich based on Krein space in this paper.Then transform problem into a quadratic opti mal value problem,systematic description and detailed theoretical proof is done.The motion model can be used in order to improve the precision of the strong tracking fil tering algorithm.While as the simple motion model can no longer satisfy the motion problem in complex environment,an interactive model algorithm is proposed.The algo rithm is combined with robust H? filtering,and the simulation results show that the algorithm is efficient,and the high efficiency of this method is proved by numerical s imulation and comparison experiment.The cooperative localization technology of multiple robot is an important part of t he path planning technology of multiple AUV.The quality of cooperative localization technology directly affects the efficiency of the whole system and the completion of t he task.In order to improve the accuracy of multi vehicle cooperative positioning,a robustness H? filtering algorithm for a nonlinear uncertain system is proposed base d on Krein spaces in this paper.The motion condition of underwater vehicle is appro ximate to the motion condition of the land multi-mobile robot.The algorithm is applie d to the collaborative localization of land multi-mobile robots,and then increase the number of robots artificially.This can intangible increases the complexity of the syst em and fully reflects the stability and high efficiency of the algorithm.It can be seen from the results that this algorithm can effectively reduce the positioning error of the multi-vehicle and greatly improve the navigation and positioning accuracy of the vehi cle system.